Chronic pain is a widespread international health problem that imposes costs of over 600 billion dollars per year. This application focuses on fibromyalgia (FM), which is characterized by persistent, widespread body pain, with significant evidence of altered brain function. One of the few effective treatments for FM is cognitive-behavioral therapy (CBT), which has been shown to reduce pain intensity and pain-related disability, potentially via reductions in catastrophizing, an important psychosocial factor that plays a crucial role in shaping individual differences in pain- related outcomes. It is likely that brain mechanisms underpin the beneficial effects of lowered catastrophizing and reduced pain in FM, and CBT-produced decrements in catastrophizing may act to normalize dysfunctional central nervous system pain processing. Our previous study in FM patients confirms that patients who are high in catastrophizing show enhanced pain-related activation in brain areas that process emotional aspects of pain (e.g., anterior insula and medial thalamus). In addition, catastrophizing was associated with altered pain-evoked functional connectivity between thalamus, anterior insula, and default mode network (DMN) structures such as medial prefrontal cortex. The DMN is a constellation of brain regions that mediate self- focused cognitive processing; our prior work has implicated altered resting-state functional connectivity between DMN and insula as a biomarker for clinical pain in FM. In our pilot study, a subset of FM patients was randomized to CBT or an active educational control condition. We found that CBT reduced catastrophizing and produced corresponding changes in brain activation and DMN connectivity. In the present proposal, we hypothesize that CBT-produced improvements in pain will be anticipated and mediated by reductions in catastrophizing and their associated effects on pain-related brain functioning. Participants with FM will be randomized to 8 weekly treatments with CBT or an education/attention control, and followed up for 6 months. Functional MRI data, including functional connectivity during both a resting and evoked deep- tissue pain state, will be collected at baseline, mid-treatment, and after the final treatment visit. We hypothesize that CBT will reduce catastrophizing early in treatment, resulting in adaptive changes in the brain's responses to an externally-applied noxious stimulus. These changes will then predict subsequent changes in patients' resting state connectivity in DMN and pain- relevant brain regions, which will underlie long-term improvements in clinical pain. Our overarching goal is to understand the neurobiological pathways by which CBT and reduced catastrophizing facilitate improvements in chronic pain; such information will help to refine biopsychosocial models of pain, identify potential non-responders early in treatment, and facilitate the enhancement of psychosocial interventions for chronic musculoskeletal pain.